Why I built Imago

I’ve spent most of this year building with AI. I’ve also spent a good part of it worried about AI. Both are true, and Imago is what I built in between.

We already ran this experiment once

We ran it with smartphones and social media. We handed a generation the feed, and only years later started counting the cost in sleep, attention and friendship. Jonathan Haidt spent a book counting it. His diagnosis in The Anxious Generation: “overprotection in the real world and underprotection in the virtual world” is why “children born after 1995 became the anxious generation.” Australia banned social media for under-16s only last December, more than a decade after the feed arrived. Singapore is writing its own safeguards into law now. We aligned the technology long after the damage was done. As Jay Kim, a pastor in Silicon Valley, put it: “We were not prepared. We started playing catch-up and we’re still trying to play catch-up.”

That line comes from a conversation I keep coming back to: Andy Crouch and Jay Kim on Practicing the Way, hosted by John Mark Comer. It opens by naming 2007, the year the iPhone launched, as a before-and-after moment the church walked straight past:

Many of us uncritically adopted technologies like the smartphone or social media without knowing just how these new tools were designed to malform our soul… we missed a chance to pastor people into an alternative future. Now we’re seeing the emergence of AI and we think right now might be a 2007 moment for artificial intelligence.

A 2007 moment. AI is the next wave of that same kind, and a more powerful one. In August, Bill Gates wrote that we are “only beginning to understand the dangers that the internet — especially social media — can pose to young people’s development,” that “AI could magnify many of these risks,” and that “we should not wait another generation to start taking them seriously.”

This is the part that stays with me: we already know how this story goes. Smartphones and social media taught us, at real cost, what happens when you figure out the harm afterwards. So when AI arrived, it was like history shouting at us in flashing lights: here it is again, the next thing. Don’t mess it up this time.

That makes this a responsibility and an opportunity. Crouch observes that pastors and communities “have more of a shot of really shaping how our communities respond to this than we had in the previous two” waves, because this time, nobody is blindly enchanted. This time the alarm is early. So I take it as a duty, and honestly as a gift: to steward this technology well, to maximise what it can give us and to guard against what we already know it can take.

Remaining human

In May, Pope Leo XIV published his first encyclical, Magnifica Humanitas, on safeguarding the human person in the time of AI. I found one of his lines hard to improve on:

In the era of artificial intelligence, when human dignity is threatened by new forms of dehumanization, ours is the pressing duty to remain profoundly human.

Remain profoundly human. What does that mean?

I don’t think it starts with knowledge. Knowledge is good, but it puffs up, and it was always secondary. When we ask whether someone is truly human to us, we don’t ask what they know. We ask: can I trust you? Do I know you? Have we sat at the same table? Being human is about connection: to feel with people, to empathise, to converse, to belong.

That is the original design. In the beginning, God made humanity to live in a constant flow of union with the Father, the Son and the Holy Spirit, and in community with one another. Knowledge was never the point of that design; love was. Every wave of technology when used unwisely pulls us out of that flow, out of embodied, trusting, present relationships, makes us a little less human. AI is the newest and strongest pull yet, and also the newest chance to choose otherwise.

Crouch ends that same conversation with a test drawn from the greatest commandment. To be human, he says, is to be “a heart, soul, mind, strength complex designed for love.” So his rule for any AI tool is simple: use it only if it grows that capacity for love in you, “and if it doesn’t, I will gently but firmly say no thank you.” I couldn’t find a better filter for what Imago should be.

Human reserved

In his August post, Bill Gates makes a proposal he calls Human Reserved: some things should be set aside for human beings the way we set aside nature reserves: “places where we could put buildings and roads, but we choose not to because the loss would be too great.”

His example is a caregiver: work that is irreplaceably human. Some things can be delegated to a machine the way you delegate a task, but they cannot be delivered by one. A robot could technically tell you that you have an incurable disease. It just shouldn’t.

Gates discusses this in an economic sense but I’d like to take it one step further. Human Reserved is also social and spiritual. There are parts of life that no productivity gain should touch: prayer, friendship, sitting with someone in grief, the conversation after a sermon.

AI has already given us so much productivity that we can afford this. We can afford to slow down. We can actually afford to be inefficient. We can afford to simplify. I say that as someone deep in this industry and genuinely excited by every new model. Today, it may be unrealistic to slow down at work or in school. But there are so many other corners of life, most of life really, where we can deliberately choose the slower, more human way. Imago is my attempt to build in one of those corners.

So why build anything with AI at all?

Because I’ve also seen what it does well. It helped me learn faster this year than I have in a long time. It’s good at helping you practise things that have a structure to them, like memorising a verse. And when it’s built carefully, it feeds curiosity instead of stifling it.

So Imago tries to keep the good and leave out the rest: less doom-scrolling, more time in the Word. Less indifference, more noticing. A tool that points past itself to serve a bigger purpose.

(Some of) What Imago does

Every feature is a stepping stone, not a destination. The summary is there to send you back to the sermon. The verse practice is there to write Scripture on your heart. The chat is there to start godly conversations, not to replace them.

What I chose not to build

The chat is deliberately not a companion. It won’t try to be your friend.

Crouch gives one practical reason: an entity that is predictable and always attuned to you “is actually terrible training for human relationships… AI doesn’t need you. It’s not going to get sick. It’s not going to require compassion. It will not grow old.” Real people need you, interrupt you, age, forgive you and are forgiven by you. That friction is the training ground of love, and a frictionless companion quietly untrains you for it.

My deeper reason is theological: AI does not have a spirit. God has a Spirit, and you have a spirit, and connection happens between spirits. A chatbot can process words about God, but it cannot know him, and you were not made to commune with it. So Imago won’t tell you God is speaking to you, and it won’t prophesy. It won’t give you a yes or no on life decisions like divorce, because those belong with God, your pastors and the people who know you. If you’re in crisis, it points you to Samaritans of Singapore on 1-767.

These aren’t just intentions. They’re written as tests, and Imago has to pass all these seven crucial areas every time I check it.

I built this as carefully as I know how. My hope is that you’ll use it just as carefully.

Imago should send you back to the things that actually form you: Scripture, prayer, and the people sitting next to you on Sunday. If it ever starts replacing those, I’ve built it wrong, and I’d want you to tell me.

If you want a place to start: after service this Sunday, open Imago and ask it one question about the sermon. Then ask someone next to you what they made of it.

That second conversation is the whole point. Because that’s the bet behind building Imago with AI at all: that technology, used carefully, can do two things. It can connect you more deeply to God, and it can send you into godly conversation with a friend, whether they’re in the church community or miles outside it. Both of those point to the same end.

Imago exists because of AI. This whole application was built with it. Two years ago I couldn’t have made it; today I could, and I’m thankful to God for that. But the technology is the means, not the end. The end is the thing it serves: spiritual growth. And spiritual growth, by everything above, doesn’t take you away from being human. Christ is the true human, the perfect image of God, which is what the name Imago points to. So becoming more like him is becoming more human, not less, and through him, more deeply connected to God.